Olives @olives@qoto.org

https://www.theguardian.com/technology/2023/aug/15/digital-grooming-crimes-online-safety-bill-nspcc
Alright, it's not images this time.

Multiple offenses can be levelled against the same person in relation to the same victim. In fact, this is not an unusual practice. We see it all the time. Perhaps, that contributes to these statistics.

The headline is extremely misleading. They mention their statistic is "in wait for the bill", which makes people think of months, or a year, however, they actually have a very wide window of many, many years.

Curiously, it seems they only created this "speaking sexually with a minor" offense in 2017 / 2018. It's hard to extrapolate off that.

Presumably, this flags anyone who speaks inappropriately with a minor. While inappropriate, that might not be someone looking to engage in more sinister activities.

In any case, it's not hard to surmise there are a certain number of crimes on the Internet.

As I've mentioned before, we don't search random homes for dead bodies, nor do we remove the doors off locks to theoretically make it easier to do so.

Also, from the looks of some of the discourse around this, they are catching predators? It seems from reports made to the police?

In one case, the victim attributed their ability to get the police to intervene in a case where someone was trying to extort them, to an online safety program they attended in school.

The criminal was later convicted for his crimes.

This begs the question as to why these people aren't advocating for more countries to run such programs.

Instead, they lobby for pieces of legislation which are disproportionate, bad for human rights, and are very likely, ineffective.

The article also uses sensationalistic language to try to make it appear as if the process of getting someone's confidence is easier than it really is. Given the framing around it, this isn't terribly surprising.

Lots of assertions about a very dubious bill (and what it "might" do) from lobbyists asking the government to "do something".

There's a random line complaining about "AI" (whatever that is supposed to mean), even though it is irrelevant to this, because got to have that in there. Again, human rights.

https://en.wikipedia.org/wiki/PA_Media
"PA Media" wrote this article. Who is that? The biggest shareholder is a very right wing company, although I presume it isn't just them.

Noting that staff for social networks mention that victims are typically of an older cohort... Hmm.

Apparently, X is aggressively attacking VPNs, because privacy and basic human rights is too much to ask, I guess. #dsa

Strong feeling of deja vu.

What have we here. An argument over whether Biden was a racist at some point.

Well, I don't know a whole lot about that, however, he's probably better than the alternatives.

I kind of feel sorry for the main character.

Does anyone remember this anime? I certainly do.

https://www.techdirt.com/2023/08/14/wisconsin-pushing-bill-that-requires-websites-to-treat-all-users-as-if-theyre-children/
Okay, I wrote an earlier version of this, but I wanted to sum it up better, so here's another post.

Some things to consider here:

1) IPs are not individuals (one might be shared by many on a mobile network).

Or someone might use a VPN. This might be for a number of reasons, including for security at a less secure hotspot.

It's very presumptuous to assume it's simply to bypass an "age restriction". There are many reasons other than this.

2) Geo IP is not always reliable, particularly at a smaller scale (i.e. cities).

3) IPs are cycled fairly frequently. How do I know someone's intent for doing so?

4) Feedback about things going wrong might not reach a platform. Often, it won't.

5) Parenting could address minors fooling around at night. It's already widely practiced.

Lawmakers here have not shown that the problem they look to address is actually a problem.

6) Accuracy is hard to measure. Exaggerated marketing materials, or artificial tests in labs, have a tendency of catching people's attention.

7) At scale, any inaccuracy is going to be a problem, and will lock out a lot of people (who may be vulnerable and / or disproportionately affected).

I've seen these systems going wrong in practice. Often, that is it, you are out.

8) If someone moves, have they "actually moved", or does a social network have to assume they're using a VPN?

Geo IP is typically made for countries (although, how reliable it is there is always unclear).

It can be hilariously bad when applied to cities.

Also, what if someone's IP changes? This happens very regularly and without their intervention.

Is this an attempt at "bypassing" a control? How would you know?

If this is just for a geoblock bypass, it still poses constitutional, jurisdictional, privacy, and other headaches. It's a mess and undermines human rights.

But, even in that case, why are lawmakers constraining how a network might be designed? Where is it the state's business to get involved?

The thing with the IPs is basic community admin knowledge.

I've known people with dozens of users who have run into the "mobile network" problem. And you expect this to hold up at Facebook scale? Hah.

Another bill from people who fundamentally don't understand the Internet, although I suppose if they did, then I suspect they wouldn't have come up with this kind of "think of the children" bill in the first place.

https://www.techdirt.com/2023/08/14/wisconsin-pushing-bill-that-requires-websites-to-treat-all-users-as-if-theyre-children/
Did you know that on a mobile network, many many people might share the same IP Address? Pretty mundane thing.

There are other ways in which IPs are unreliable and don't mean a specific person. Again, lawmakers making very shallow assumptions here...

This is the same mistake Elon made and it bit him too.

Now, this might sound radical, but usually parenting prevents minors just fooling around all night (although gotta sleep too). Nanny state much?

How do you measure "accuracy"? Is it exaggerated marketing materials? Is it artificial tests in the lab?

Also, at scale, that is still a nightmare, and can effectively lock a lot of people out (who might quite vulnerable and / or disproportionately affected).

I've seen these systems going wrong in practice. Often, that is it, you are out.